Comparison of dry separation processes for

In addition to air classification, dry separation of cereal grain components can be achieved by sieving or applying electrostatic separation. In this section, these two methods are shortly presented, and their advantages and disadvantages when compared with air classification are evaluated in Table 5.

Sieving is a traditional dry fractionation method, employed for the separation of particles based on their sizes between one or several meshes of decreasing size. Sieving is considered the method with the most potential for dry particle separation when the particles exhibit sizes larger than 100–200 µm. Rhodes (2008) reviewed that normal or air-jet sieve sizes larger than 45 and 20 µm, respectively, provide reliable results in sieving-based particle size analysis. As reviewed by Furchner and Zampini (2012), sieving is technologically easier to carry out than air classification and requires less energy. However, the main limitations in sieving include clogging or blinding of the sieves, especially with high-fat and ultra-fine materials, and the separation mechanism being limited to only separating according to the size and shape differences of the particles. As proteins are found in the finest-sized particles and fibrous plant cell wall materials usually remain in larger cellular integrities, size-based separation and the enrichment of protein and fibre in the fine and coarse fractions, respectively, may be achieved via sieving (Schutyser and van der Goot 2011).

The fractionation of components in an electrostatic separator takes place as a result of differing electrical charges that the components receive during a charging step. The charges obtained are component and species specific and are mainly affected by the surface properties of the materials, such as surface chemistry, electrical conductivity and dielectric properties (Flynn et al. 2019;

Mazumder et al. 2006; Németh et al. 2003). Plant material fractionation by electrostatic separation is based on either conductive induction, ion or corona charging, or tribo-charging. As reviewed by Barakat and Mayer-Laigle (2017), in a parallel-plate tribo-electrostatic separator particles are tribo-charged as a result of collisions of the particles with both each other and the equipment walls inside the charging pipe, and fractionation takes place in a separation chamber where the positively and negatively charged particles adhere to the high voltage electrodes of opposite charges and uncharged particles separate by gravity. In conductive belt or drum separators, particles with different conductivities are separated from each other (Flynn et al. 2019; Higashiyama and Asano 1998).

Table 5. Comparison of dry fractionation approaches.

Property Air classification Sieving Electrostatic separation

1–200 µm Preferably larger than 100–200 µm; normal

cereal grains have mostly concentrated on DF enrichment. DF enrichment by electrostatic separation has been achieved for defatted and pin disc-milled rice bran (from 32 to 52% DF with a mass yield of 44% in a one-step process and from 32 to 67% DF with a mass yield of 21% in a two-step process) (Wang et al. 2016) and pin disc-milled wheat bran (from 23 to 31% DF with 16% mass yield) (J. Wang et al. 2015b), whereas the same authors revealed more successful separation by using only a simplier air jet sieving (68% DF with 22% mass yield for rice bran and 31% DF with 45% mass yield for wheat bran).

Protein content was not considerably changed in any of the fractions (J. Wang et al. 2015b; Wang et al. 2016). For wheat bran, further improvements were obtained by adding an air jet sieving step after the electrostatic separation, which resulted in arabinoxylan content of 43% with an 8% mass yield (J. Wang et al. 2015b). Similar data supporting better DF enrichment by sieving than by air classification has also been reported for barley (Knuckles and Chiu 1995).

On the contrary, Wu et al. (1994) observed for non-defatted and defatted high-β-glucan barley variety samples high-β-glucan enrichment from 17.4–19.6 to 24.6–

28.3% in a sieving fraction with approximately 5% mass yield, whereas higher β-glucan contents and mass yields were obtained by air classification. In regard to oats, the sieving of a defatted and pin disc-milled oat bran allowed the production of several fractions with β-glucan content of 19.9–20.7%

compared with 11.1% in the raw material (Wu and Doehlert 2002). Higher contents were reached in electrostatic separation by Sibakov et al. (2014) who demonstrated that defatting enhanced β-glucan separation from ultra-finely milled oat bran and the contents were increased from the original 21.3–35.0 to 26.3–42.2% and 31.2–48.4% with mass yields of 53.1–46.2% and 27.3–

23.1% by applying one- and two-step electrostatic separation processes, respectively. On the contrary, air classification of the same materials allowed enrichment from 21.3 to 35.0% and, further, from 35.0 to 40.3% with mass yields of 54.5% and 84.6%, suggesting that from a high β-glucan material (35.0% β-glucan), more enrichment is achieved by electrostatic separation (up to 48.4%) than by air classification (up to 40.3%), whereas separation efficiencies were higher in air classification due to higher mass yields. On the contrary, in another approach that aimed at histological aleurone enrichment, several steps of electrostatic separations performed for impact-milled rye and wheat brans decreased the total DF content but increased the ratio of SDF and IDF (Nordlund et al. 2012).

Research on successful cereal protein enrichment by electrostatic separation is lacking in literature. For starch-rich pulses, air classification has proven more efficient than electrostatic separation, whereas electrostatic separation is more applicable for oil-rich pulses and oilseeds, such as rapeseed and sunflower seed press cakes (Laguna et al. 2018; Pelgrom et al.

2015). In regard to sieving, Prakash and Ramanatham (1994) were able to reach protein content of 18–20% in sieving fractions produced from a defatted and roller-milled rice bran that initially had 17% protein content, and Jayadeep

et al. (2009) produced a protein-enriched fraction (max. 19.3% protein, <5%

mass yield) from a defatted and crushed rice bran (14.5% protein). Sumner et al. (1985) separated different pearling fractions from barley and reached a maximum protein content of 27.0% with 17% mass yield from the hulless non-waxy barley (19.9% protein). Sieving a high-protein, high-beta-glucan barley variety (23.3% protein) resulted in a fraction <64 µm with 26.6% protein with 44% mass yield, and protein enrichment was further improved by defatting the relatively high-fat (5.7% fat) barley material, that allowed to reach protein content of 29.1% with 15% mass yield in particles <43 µm, whereas higher PSE was obtained by air classification (Wu et al. 1994). Protein enrichment to 21.8% in an aleurone fraction produced from wheat bran (15.2% protein;

Hemery et al. 2009) has been reached using different dry processes based on descriptions in Bohm et al. (2003) and Bohm and Kratzer (2005), and for rye, the fractionation of rye aleurone from whole rye flour by sieving resulted in protein enrichment from 11.4 to 17.6% (Glitsø and Bach Knudsen 1999). For defatted and pin disc-milled oat bran, the protein content has been increased from 28.6% to over 32% in multiple sieving fractions with low mass yields (Wu and Doehlert 2002).

2.2.4 FACTORS AFFECTING THE EFFICACY OF DRY